In 3GPP (3rd Generation Partnership Project), the packet-switched communication systems HSPA (High Speed Packet Access) and LTE (Long Term Evolution) have been specified for wireless transmission of data packets between user terminals and base stations in a cellular/mobile network. In this description, the term “base station” is used to generally represent any network node capable of wireless communication with a user terminal.
LTE systems generally use OFDM (Orthogonal Frequency Division Multiplexing) involving multiple narrow-band sub-carriers which are further divided into time slots to form a so-called “time-frequency grid” where each frequency/timeslot combination is referred to as a “Resource Element RE”. In LTE, multiple antennas can also be employed in both user terminals and base stations for obtaining parallel and spatially multiplexed data streams, e.g. according to MIMO (Multiple Input Multiple Output), which is well-known in the art. Other wireless communication systems relevant for the following description include WCDMA (Wideband Code Division Multiple Access), WiMAX, UMB (Ultra Mobile Broadband), GPRS (General Packet Radio Service) and GSM (Global System for Mobile communications).
A base station of a cell in a wireless network may transmit data and control information in a physical downlink channel to a user terminal or “UE” (User Equipment), and a user terminal may likewise transmit data and control information in a physical uplink channel in the opposite direction to the base station. In this description, a physical downlink or uplink channel is generally referred to as a wireless link between a sending node and a receiving node. Further, the terms “sending node” and “receiving node” are used here merely to imply the direction of the wireless link considered, although these nodes can of course both receive and send data and messages in an ongoing communication. Further, the term “Resource Element RE” is used in this description to generally represent a signal bearer element that can carry a signal over a wireless link, without limitation to any transmission technology such as LTE. For example, an RE can incorporate a specific code and timeslot in a system using CDMA (Code Division Multiple Access), or a specific frequency and timeslot in a system using TDMA (Time Division Multiple Access), and so forth.
When two nodes in a cell communicate over a wireless link that is configured according to various link parameters, one or more such link parameters can be adapted to the current state of the link on a dynamic basis, often referred to as link adaptation. Such link parameters may include transmission power, modulation schemes, encoding schemes, multiplexing schemes, and the number of parallel data streams when multiple antennas are used, the latter link parameter being called “transmission rank”. Link adaptation is used to generally optimise transmission in order to increase capacity and data throughput in the network. Further, link adaptation can be employed for the uplink and the downlink independently, if applicable, since the current state of the uplink and downlink can be very different, e.g. due to different interference and when frequency and/or time are widely separated for uplink and downlink transmissions between the two nodes.
To support link adaptation during an ongoing communication between a sending node and a receiving node, either on the uplink or downlink, the receiving node is often required to measure certain link parameters and report recommended link parameters to the sending node, such as a recommended transmission rank and/or a recommended precoder matrix. Also, the quality of the received signal is often measured, typically in terms of a Signal to Interference and Noise Ratio SINR, e.g. separately for different parallel data streams, assuming that the recommended link parameters are used by the sending node. Based on the recommended link parameters and measured SINR value(s), the receiving node estimates so-called “Channel Quality Indicators” CQIs, e.g. one CQI for each coded data block (codeword), that are used together with the link parameters to indicate the current state of the link, which is reported back to the sending node. In this description, a reported CQI or the equivalent and/or recommended link parameters will be called a “link state report” for short. The sending node can then adapt one or more link parameters depending on the received link state report. When the sending node is a base station using packet switching for downlink transmissions, the reported CQIs may also be used for packet scheduling decisions.
Typically, specific known reference symbols RS are regularly transmitted over a wireless link according to a predetermined scheme to support the above link quality estimation, such that the receiving node is able to detect noise and interference more easily without having to decode the received signal. In an OFDM-based LTE system, these RSs are transmitted from base stations in predetermined REs in the time-frequency grid as known by the receiving terminal.
In general, a received signal “r” in an RE is basically comprised of transmitted symbols “s” as well as noise and interference “n”. Thus:r=Hs+n  (1)
Generally, r, s and n are vectors and H is a matrix, where “H” represents the channel response which can be derived from a channel estimator in the receiver. However, the noise and interference of a signal in an RE display different characteristics depending on whether the RE contains payload data, control signalling or an RS, as the interference mix hitting the different types of REs may typically have different transmission power and spatial characteristics, e.g. due to time and/or frequency synchronization in neighboring cells. The interference/noise “I” in these different signal types may be characterised in terms of second order statistics that can be obtained by frequently measuring the signals over time, although “I” can be characterised in other ways as well.
If an RE contains an RS signal received by a user terminal, the terminal is able to estimate the interference/noise n=I(RS) of the RS signal since s are known symbols in this case and H is given by the channel estimator. If the RE contains data scheduled for the terminal, the interference/noise n=I(data) can also be estimated once the data symbols have been detected (i.e. decoded) by the terminal, s thereby being known at that point. Similarly, the interference/noise of an RE with control signalling, n=I(control), can be estimated if the control symbols can be detected.
In order to obtain proper link quality estimation and to determine an accurate CQI and/or link parameter recommendation for a link, the receiving node needs sufficient statistics from measuring signals transmitted on the link. Furthermore, the characteristics of inter-cell interference may be significantly different depending on what signal type is causing the interference from neighbouring cells, i.e. RS signals, data signals or control signals. If payload data is transmitted over the link to be estimated, the receiving node should preferably measure the interference I(data) that hits the data signals. However, the measurements would then be limited to REs that contain data scheduled for the user terminal involved, which may be too scarce such that the statistic basis for determining the CQI is insufficient. Moreover, the data symbols must be detected and decoded, and possibly also re-encoded, before the interference I(data) can be properly estimated, which may impose substantial costs and/or unacceptable delays due to the data processing.
Alternatively or additionally, the receiving node can measure the interference I(RS) for REs containing an RS which may occur more frequently than the REs containing scheduled data. Measuring I(RS) is also generally more reliable since the RS is always known to the receiving node. However, the interference that hits RS signals may be significantly different from that hitting the data signals, e.g. with respect to statistics. Therefore, a CQI and/or link parameter recommendation determined from I(RS) measurements may not be representative for a link with payload data transmission. As a result, the link adaptation at the sending node may not be optimal for data due to either too optimistic or too pessimistic CQI and/or link parameter recommendation from the receiving node. Hence, if the measured I(RS) is significantly greater than the actual I(data), the CQI and/or link parameter recommendation will be based on an overestimated interference (or underestimated SINR) and therefore unduly pessimistic, and vice versa.
For example, when MIMO is employed in an LTE system, the RE holding an RS from one antenna at the sending node must be empty for a neighbouring antenna, which substantially limits the number of REs available for RS transmissions. As a result, the interference that hits REs containing an RS will largely come from RS transmissions in other cells due to reuse of the RS transmission pattern. As mentioned above, RSs are always transmitted from base stations according to a predetermined scheme and at a relatively high fixed power in order to be received by any terminal in the cell, whereas payload data is only transmitted when scheduled for a specific terminal. Thus, in a situation with low data traffic and/or low transmission power for data signals, I(data) is generally lower than I(RS).
Furthermore, control signals are often transmitted with greater power than data signals, due to different power regulation. Therefore, the interference measured for an RE affected by control signal interference may be different from that of an RE affected by data signal interference.
Hence, it is often difficult to obtain accurate estimates of the inter-cell interference that hits data transmissions, in particular if the interference measurements are performed on RS transmissions, as explained above. Inaccurate estimates of the SINR may thus result in misleading CQIs and non-optimal link parameter recommendations such as transmission rank. A consequence for MIMO systems is that an underestimated SINR may result in a too pessimistic transmission rank when the used link can actually support a transmission rank greater than the recommended one. Both of these issues may well result in reduced throughput. On the other hand, if the SINR is overestimated, the link may not be able to support any recommended CQIs (including a recommended Modulation and Coding Scheme MCS) and transmission rank, resulting in excessive decoding errors and thereby reduced throughput also in this case.
However, the base station may monitor so-called “ACK/NACK signalling” from the terminal for received data blocks, and detect if a Block Error Rate BLER or the like is below or above a predetermined target value. From this information, the base station can decide to use a more offensive or defensive MCS than recommended by the terminal. However, if the base station selects a transmission rank different from the recommended one, the reported CQI will be largely irrelevant since, in most cases, it relates directly to the transmission rank. Consequently, the base station would not have a proper basis for selecting the MCS and other link parameters for the different data streams.
It is thus generally a problem that, in a communication with dynamic link adaptation, a signal sending node may receive inaccurate link quality estimations and/or link parameter recommendations from a signal receiving node, such that the used link parameters are not optimal or appropriate for the actual link used in the communication.